ETSI's Experiential Networked Intelligence (ENI) Industry Specification Group (ISG) has just released a flurry of documents related to its rather curious remit, which spans policy control, service deployment, resource management, monitoring, analysis and prediction. (See ETSI's Experiential Networked Intelligence Group Releases Docs .)

Curious because of the unusual word "experiential," by which it means "observe and learn from the experience an operator has in managing the network to improve its understanding of the operator experience, over time." Or what we might normally call continuous improvement, as exemplified by the Japanese concept of kaizen that arose in the 1980s.

Having originally published a rather vague white paper in October 2017, ENI has followed up with some very detailed use cases in the areas of infrastructure management (maintenance and planning), network operations (run-time stuff), service orchestration and management (order to activation, SLA management, etc.), and assurance (monitoring and prediction). Examples include VoLTE optimization, network slicing for IoT security and SD-WAN management.

It has also published a requirements document covering service and network aspects (service orchestration and management, network planning and deployment, network optimization, resilience and reliability, security and privacy), functional aspects (data collection and analysis, policy management, data learning, interworking), and other aspects (performance, operational, regulatory, policy). This rather dry document will inform the system architecture specification that is due to be published next year.

Dr. John Strassner of Huawei Technologies has written the Context-Aware Policy Management Gap Analysis that draws on the work he is also doing for the MEF's Policy-Driven Orchestration (PDO) initiative, although "it is likely that ENI will need additional RPs [interface Reference Points]." In terms of the gap analysis, the MEF's PDO fully meets all of ENI's requirements, except the use of AI to construct policies -- however, "this is not hard to do because the MEF PDO information model uses software patterns," says Strassner. The IETF's Simplified Use of Policy Abstractions (SUPA) fully or partially meets most of ENI's requirements, whereas the TM Forum's SID satisfies hardly any.

The ENI ISG's Proof of Concepts Framework document is of little interest in itself: However, the PoCs that must adhere to this blueprint will be critical to the success of ENI in the real world, in our view. The first PoC will be championed by China Telecom, which was a major contributor to the original white paper -- five of the 26 contributing authors were from the Chinese operator -- as well as the new use case and requirements documents.

We understand that other PoCs are in the pipeline. We wouldn't be surprised to see Verizon champion a future PoC; it is already planning to contribute a use case and some new requirements to the next version of the ISG's documents, and the rapporteur of the pending architecture spec is a Verizon representative.

ENI seems like an interesting initiative and the new use case document has provided some real-world tangibility to what previously seemed somewhat ethereal. The key will be getting other operators besides China Telecom actively involved: Telefónica, TIM (Telecom Italia) and Vodafone are all members of the group, but have not yet put their name to the specifications or announced a PoC.

It will also be key to get more vendor participation outside of Huawei, which contributed almost half of the authors of the original white paper and the new use case and requirements documents. Intel and Samsung have been involved in the ISG: It's time for some of the other industry heavyweights to lend a hand.

— James Crawshaw, Senior Analyst, Heavy Reading